SMART issues with Western Digital WD Red WD50EFRX

Hello OMV community,

I'd like to report an issue which I have somehow "fixed" by using a workaround. First of all some information about my hardware. My server is a Fujitsu PRIMERGY MX130 S2 and I have recently bought 3 WD Red 5TB (WD50EFRX). I have set up a RAID 5 array with those 3 disks using the GUI of the latest OpenMediaVault 2. Everything went fine so far and I didn't have any problems setting up the system.

I have set up S.M.A.R.T. monitoring for all 3 drives and at some point I started to receive some emails that "ATA errors" have occured (for all of the 3 disks):

Zitat

This email was generated by the smartd daemon running on:

host name: HomeNAS
DNS domain: [Unknown]
NIS domain: (none)

The following warning/error was logged by the smartd daemon:Device: /dev/disk/by-id/ata-WDC_WD50EFRX-XXXXXXX_WD-XXXXXXXXXXXX [SAT], ATA error count increased from 4 to 5

For details see host's SYSLOG.

You can also use the smartctl utility for further investigation.
The original email about this issue was sent at Sat Mar 12 19:20:06 2016 CET
Another email message will be sent in 24 hours if the problem persists.

Alles anzeigen

So I immediately checked the SMART values using "smartctl -ia /dev/sdX" and found multiple entries about ATA errors at the buttom of the output (the SMART values itself were good):

It seems like there have been multiple SMART requests at almost the same time. After noticing those ATA errors on all devices I decided to run a SMART short test on my drives (smartctl -t short /dev/sdX). Unfortunately the test did not finish in 2 minutes like it was supposed to be. The process just hang at 90% all the time on all drives. What I did then was installing the backports kernel because I thought maybe the "outdated" kernel causes this behavious.

OMV 2 with backports kernel

root@HomeNAS:~# uname -a
Linux HomeNAS 3.16.0-0.bpo.4-amd64 #1 SMP Debian 3.16.7-ckt20-1+deb8u3~bpo70+1 (2016-01-19) x86_64 GNU/Linux

While running the system and reading SMART values new ATA errors still occured. Additionally the SMART self test stopped at 90% and didnt finish (I had to manually abort it). I was really annoyed by that and decided to put the disks one by one into my desktop PC and run an UBUNTU live CD (14.04). The self test finished properly for all 3 disks. The next step was to put the drives back into the server and run the UBUNTU live CD there to rule out any hardware issues. Fortunatly the smart self test worked fine running ubuntu on my server. I was really happy to ruled out any hardware issue or incompatability of my board/controller.

Now I just had to find out how I could get OpenMediaVault to work with my server and hard drives. Therefore I decided to install the OMV 3 beta on an USB drive, update to the latest version and give it another try.

OMV 3.0.13

root@HomeNAS:~# uname -a
Linux HomeNAS 3.16.0-4-amd64 #1 SMP Debian 3.16.7-ckt20-1+deb8u4 (2016-02-29) x86_64 GNU/Linux

I couldn't have been happier when I saw that the smart self test finished without any errors. Now I hope that the ATA errors just occured because of some incompatability with the software/kernel and that it did not damage my (expensive) drives. I have just tried to document my experience. Maybe some of the devs/mods/pros know what the problem was or maybe this could help to fix/detect any further issues. If you need any logs or complete SMART outputs, I'll post them here.

Zitat von ryecoaaron

I would reinstall OMV 2.x and install the backports 3.16 kernel. This is the same kernel as the standard OMV 3.x kernel. There is a button to install in omv-extras. OMV 3.x is not ready yet.

I don't mean to be rude, but I have exactly done that as you can read in my first post. But unfortunately this did not change anything.

It may be true that the kernels are the same but nevertheless it is fact that self tests work now

Zitat von subzero79

ImO self tests are useless. You can have 1000 bad sectors uncorrectable self tests will still pass. Smart attributes values are the ones to keep track and observe.

That may be true, but I just cared about the fact that something prevented the self test from finishing